Mail Archives: djgpp/2000/02/17/22:23:50
> DJGPP already hides LOTS of these setup problems. Look at DJGPP.ENV,
> and you will understand how much *more* complex the setup would be
> without that automagic. Nothing, including DJGPP's origins, can
> prevent somebody from creating an installer that hides what's left
> (setting PATH and DJGPP variables).
Well, except for the frequent updates! <big grin>
> Some people, like Shawn Hargreaves, have an exactly opposite view.
> YMMV.
An opinion. We are all entitled to it. See one of the other posts (about
DJGPP's nearptr) for an argument in Watcom's defense.
> FWIW, the DJGPP's way of doing things is explained in this URL:
>
> http://www.delorie.com/djgpp/doc/eli-m17n99.html
Already read it, granted it was a while ago. Wait a minute... eli-m*...Eli
Zaretskii... I thought I recognized your name. I am beginning to doubt my
abilities in arguing about DJGPP with you. =) Although it's cool to be bumping
heads with you.
I suppose it's just that I've taken the time to understand Watcom's method.
> You shouldn't need PGCC anymore, since GCC 2.95.x and later produces
> code that is as good as PGCC of yore (and latest versions of PGCC
> aren't stable enough).
>
> Btw, comparison shows that GCC now produces code that's better than
> Watcom's, see the compiler comparison page cited in the FAQ.
Watcom hasn't been updated (seriously updated, that is) for quite some time.
I've been ignoring the fact that Watcom is EOL'd in this discussion. But see,
that is a true point in describing what makes GCC a good compiler. Telling us
that Quake was written in it, and that it's free... well, that's not enough for
me.
AndrewJ
- Raw text -